- Haskell-Cafe - Haskell.org

memory leak when using "forever"
by Zoran Bosnjak 04 Nov '16

04 Nov '16

Hello all, can you please explain why does this simple program leak memory. But, if I replace loop2 with loop1 (that is: without using "forever"), then it does not leak. Is this a problem in "forever" implementation or am I misusing this function? In the sources of base: http://hackage.haskell.org/package/base-4.9.0.0/docs/src/Control.Monad.html… ... there is some mention of memory leak prevention, but it looks like something is not right in this case. import Control.Concurrent import Control.Monad import Control.Monad.Trans import Control.Monad.Trans.Reader import Control.Monad.Trans.State main :: IO () main = do --let loop1 = (liftIO $ threadDelay 1) >> loop1 let loop2 = forever (liftIO $ threadDelay 1) _ <- runStateT (runReaderT loop2 'a') 'b' return () regards, Zoran

2 3

simple diagrams about GHC development flow
by Takenobu Tani 04 Nov '16

04 Nov '16

Hi cafe, For myself and new contributors, I drew overview diagrams about GHC development flow. GHC development flow http://takenobu-hs.github.io/downloads/ghc_development_flow.pdf https://github.com/takenobu-hs/ghc-development-flow Let's enjoy :) Regards, Takenobu

1 0

Re: [Haskell-cafe] programming style...and type classes...
by Olaf Klinke 04 Nov '16

04 Nov '16

Mark, Haskell typeclasses are a wonderful thing to have. I dare say that typeclasses are what makes haskell code modular and re-usable, because you can develop an algorithm "aginst the class interface" and users only need to supply a class instance for their own data type in order to use your algorithm. What the wiki warns about is misuse or over-use of typeclasses. Typeclasses express properties that different types have in common, and in that respect they are somewhat similar to OO-classes. However, there is no hierarchy of inheritance. It is true that typeclasses, particularly the more advanced ones with e.g. multiple parameters or functional dependencies, sometimes get into your way. One example is ambiguity arising from multi-parameter type classes: Suppose you have a two-parameter class {-# LANGUAGE MultiParamTypeClasses,FunctionalDependencies #-} class Foo a b where foo :: a -> b bar :: b -> (a,[Int]) Now you write a function using only bar: f :: (Foo a b) => b -> Int f = length.snd.bar The compiler will complain that it does not know which Foo instance to use, because while bar itself mentions both class parameters a and b, snd.bar has type b -> [Int] and hence lost the information about the first class parameter. Adding a functional dependency b -> a helps in this case. But the promise that for every b there is only one a such that Foo a b holds may not be what you want. So what went wrong? Observe that bar :: b -> (a,[Int]) is equivalent to two functions bar1 :: b -> a bar2 :: b -> [Int] and bar2 probably did not belong into the class Foo. We've been cramming to much into one typeclass and should therefore split Foo into two classes. Olaf

1 0

Package metadata format - notes from discussion
by Harendra Kumar 04 Nov '16

04 Nov '16

For reference, I have organized the major points of the discussion that happened earlier on haskell-cafe regarding package metadata format (cabal vs yaml, vs toml, vs hs). Take a look at this: https://github.com/harendra-kumar/package-metadata In my opinion, in the short term it might help if the cabal format is improved, the parser is modularized and detached from the cabal tool itself so that it can be used independently by any other tools wishing to do so. For long term possibility of change of format, I do not intend to pursue or investigate it further at this point of time but others who are interested are free to do so. -harendra

6 6

better error messages for YAML
by Tobias Dammers 04 Nov '16

04 Nov '16

Hello cafe, I have a bit of a conundrum here. The situation is such that I'm using the YAML package to parse YAML configuration files; some of the values in the file format require further parsing (using my own ginger library), and all that works fine. One problem though; since this is aimed at end users, I want to preset error messages that pinpoint the exact error location in case the ginger parse fails. However, the ginger parser doesn't get to see the entire YAML source, because it runs on just an individual string value inside the YAML, and the YAML parser itself doesn't report source positions either. The unfortunate result is that while I get "technically correct" parser errors from Ginger, the source positions they report are relative to the string value, i.e., all errors are reported as being on line 1 of an unnamed source file. The relevant code is here: https://github.com/tdammers/sprinkles/blob/master/src/Web/Sprinkles/Replace… The only solution I can think of right now involves writing my own YAML parser, and probably also an Aeson drop-in replacement; I don't really want to do that though. Is there any other way to get what I want? -- Tobias Dammers - tdammers(a)gmail.com

3 2

Creating string resources
by Zachary Kessin 03 Nov '16

03 Nov '16

I have an app with a UX in GHCJS and I want to have way for a non developer to edit the text, but then have it compiled into our code base. One obvious way would be just to have haskell modules that just has the strings in it as constants. But I thought to ask has anyone else come up with a solution to this? -- Zach Kessin SquareTarget <http://squaretarget.rocks?utm_source=email-sig> Twitter: @zkessin <https://twitter.com/zkessin> Skype: zachkessin ᐧ

3 2

Re: [Haskell-cafe] Suggestion for Data Structure regarding Timetabling and Time Intervals
by dominic＠steinitz.org 03 Nov '16

03 Nov '16

I don’t know if this is relevant but for manipulating time intervals, I have found the intervals package very useful. > Message: 5 > Date: Wed, 2 Nov 2016 21:26:32 +0100 > From: Olaf Klinke <olf(a)aatal-apotheke.de> > To: list(a)mroth.net > Cc: haskell-cafe(a)haskell.org > Subject: Re: [Haskell-cafe] Suggestion for Data Structure regarding > Timetabling and Time Intervals > Message-ID: <63753058-716B-4D2B-8D8F-4B0F77732C2A(a)aatal-apotheke.de> > Content-Type: text/plain; charset=us-ascii > > Dear Michael, > > if you can be sure your intervals will never overlap, why not use a binary search tree, such as Data.Set from containers? You could make a data structure distinguishing major and minor intervals such as: > > type Time = -- your favourite totally ordered time type > data Major = Major { > fromTime :: Time, > untilTime :: Time} deriving (Eq) > data Minor = Minor { > after :: Time > before :: Time} deriving (Eq) > type TimeInterval = Either Minor Major > > instance Ord (Either Minor Major) where > Left a <= Left b = before a <= after b > Right a <= Right b = untilTime a <= fromTime b > Left a <= Right b = before a <= fromTime b > Right a <= Left b = untilTime a <= after b > > If I understood your intentions correctly, inserting a minor interval between a :: Major and b :: Major where a < b would mean to insert > > Minor {after = untilTime a, before = fromTime b} > > It might also be handy to define a new type (isomorphic to the product (a,b)) > > Anno a b = Anno a b > > whose Ord instance uses the Ord instance of b only, so that > > Anno "Some annotation" (Major t0 t1) > > is an interval bearing some annotation that can be ordered just like the Major type. > > If your intervals _do_ overlap, then two intervals can be in six different configurations (instead of LT, GT and EQ), and several tree structures exist that support some operations more or less efficiently. Bioinformatics and geometry has developed good data structures and algorithms to deal with efficient overlap searches. > > Olaf > > Dominic Steinitz dominic(a)steinitz.org http://idontgetoutmuch.wordpress.com

1 0

Using unboxed integers
by Richard A. O'Keefe 03 Nov '16

03 Nov '16

I have x, y :: Int# I expect that x <=# y :: Bool but ghc 7.8 says Couldn't match expected type âBoolâ with actual type âInt#â In the expression: x <=# y In an equation for âbâ: b = x <=# y I have MagicHash and I've imported the right modules so that <=# is recognised as *something*, but at this point I'm baffled. I'm also not clear on how I can show an Int#.

3 3

Any students for master project in The Netherlands? (near Eindhoven)
by Damian Nadales 03 Nov '16

03 Nov '16

Hi, In my current team, we're looking for Master students to work in a project about checking RAML specifications automatically. We would like to develop this in the functional paradigm, but all the students we got so far have no FP exposure whatsoever. So if you know of somebody that might be looking for a master project and want to know more details please do let me know (maybe Alejandro and Jurriën have some ideas ;)). Best, Damian.

1 0

Call for participation: BayHac 2017 survey
by danburton.email＠gmail.com 03 Nov '16

03 Nov '16

I've invited you to fill out the following form: BayHac 2017 participation survey To fill it out, visit: https://docs.google.com/forms/d/e/1FAIpQLSdaHON7_RFSwR-2aAHpR8bJ6IUvpQ7Usul… BayHac is a Haskell hackathon weekend event for the San Francisco Bay Area and Silicon Valley. It will be a weekend of learning, hacking, and connecting with fellow Haskell enthusiasts. The proposed outline for 2017 is: Friday: Haskell in Production & getting started with Haskell Saturday: Get involved! Highlights of various open source Haskell projects. Sunday: Follow along: Abstract concepts taught via Haskell code snippets. As we solidify the location and date (tentatively March or April 2017), Haskellers of all varieties and skill levels are invited to participate to the degree that they are willing and able. If you are interested in attending BayHac 2017, kindly fill out this survey to help us calibrate and organize our preparations. Google Forms: Create and analyze surveys.

1 0